Ozonoff et al. (1987) conducted a symptom prevalence survey in a neighborhood assumed to be exposed to airborne hazardous wastes. The study population included households within 400 meters of a hazardous-waste site in Lowell, Massachusetts; the unexposed controls were those households within a radius of 800 to 1200 meters from the site. Linear distance of each residence to the center of the site was determined, thereby providing a further breakdown of potential exposure. The exposure surrogate was distance from the hazardous-waste site. In contrast to the aforementioned studies by Baker et al. (1988) and Lipscomb et al. (in press), Ozonoff et al. concluded that the study “raised the possibility that exposure to relatively low levels of airborne chemicals may have increased the prevalence of respiratory and constitutional symptoms in adults in the affected neighborhood ” (Ozonoff et al., 1987, p. 596). They noted that the most serious potential problem in the study was recall bias—special importance being given by respondents to particular symptoms. Careful evaluation of the potential for recall bias indicated that six symptoms exhibited a “biological gradient” (a dose-response relationship) and recall bias does not account for the study findings.
It is outside the scope of this chapter to review each of the last three studies described (Stringfellow, McColl, and Lowell) in more detail, but it is important to point out that their exposure assessments were very sparse. That limits the confidence in a positive association between the exposed populations and subjective health outcomes. Absence of an association is equally problematic, given the lack of individual exposure data, information derived from microenvironmental monitoring, or indirect methods based on modeling. The potential for misclassification in these studies seems to be particularly high.
The problem of chemical identification and false linkages is even more intractable when we consider findings based on subjective reporting. It will be difficult to resolve these differences entirely without an improved understanding of the nature and scope of exposure.
Like Ozonoff et al. (1987), who studied distance from the waste site, Hertzman et al. (1987) used distance as their surrogate for exposure in investigations of adverse health effects associated with the Upper Ottawa Street Landfill Site in Hamilton, Ontario. In addition, they carried out a prospective morbidity study of workers as a hy-